Google brings Lyria 3.5 AI music generation to Gemini

The difference is what is not attached to it. Suno and Udio are both defending copyright suits from labels, artists and collecting societies. Google is entering the same market without that weight.

What the model does

Google calls Lyria 3.5 its best-sounding music generation model, citing more expressive vocals and richer arrangements than earlier versions.

It builds complete songs rather than looping a short melodic idea. Tracks come with verses, choruses and bridges, in high-fidelity stereo, at variable lengths up to three minutes. Users pick genres, vocal styles and templates, and the model generates a cohesive track matching a requested duration.

The model first appeared in Google Flow Music on 29 July. The September release pushes it to a far larger audience through Gemini.

Why the licensing position matters

The AI music market has been shaped so far by two companies fighting on multiple legal fronts at once. That has constrained what they can do publicly and how labels deal with them.

A well-resourced third entrant changes the shape of that market. Google can absorb legal risk that a startup cannot, and it arrives without the existing judgments and filings that Suno now carries.

For producers, this means the tools are no longer a two-horse race, and the terms on which AI music gets made are about to be negotiated by a company with considerably more leverage. AI already generates more new music than humans, and Google entering at this scale accelerates that.

Google has not detailed what Lyria 3.5 was trained on in the material accompanying the Gemini release. A cleaner legal position today is not the same as a disclosed one.

The engineers debating whether AI replaces producers keep returning to the same point. What matters is not the quality of the output but whether anyone can say where it came from.

Lyria 3.5 is available now to all Gemini users. It marks another step in how AI has reshaped electronic music over the past two years.